The pulverizing system is an indispensable auxiliary system for coal-fired power plant units. Its safe and economical operation directly affects the economical efficiency and safety of the operation of the power plant. Among them, steel ball mills have been widely used in thermal power plants as an important part of the medium storage pulverizing system, and have the advantages of reliable operation, simple maintenance, strong adaptability, and low maintenance cost. However, ball mills are multivariable, nonlinear, strongly coupled, and large delay objects, coupled with changes in raw coal moisture, coal quality (volatiles, ash, grindability, etc.), and wear of steel balls, lining plates, etc., which make ball mills work. Due to the complex nature of the conditions, it is difficult to monitor and control the ball mill. At present, most ball mills in China are controlled manually by operating personnel. Long-term manual operation not only causes ball mills to stop grinding, cut coal, run coal, and over-temperature events, but also prevents the system from operating under optimal conditions for a long period of time. Under the premise of safe operation of the ball mill, reducing coal unit consumption, improving economic efficiency and automatic operation input rate of the unit become a key to the control of the power plant.
First, the establishment of a mathematical model of the ball mill 1.1 Test Method Modeling Test method modeling is established through the relationship between the input and output signals, for example, the input signal can be used as a step signal or pulse signal to get the system's transfer function. At present, in practice, the differential pressure at the inlet and outlet of the coal mill is generally used to characterize the material level.
1.2 Mechanism Method Modeling Mechanism method modeling is to write various balance equations based on the mechanism of system change: material balance equation, energy balance equation, momentum balance equation, phase balance equation, and reflection of fluid flow, heat transfer, mass transfer, and chemical reaction. The basic equation of the characteristic equation, from which to obtain the desired mathematical model.
Second, the ball mill powder system commonly used control method 2.1 Conventional proportional integral differential (PlD) control ball mill system is a three-input three-output control object, coal supply, hot air door opening and recirculation (cold) throttle opening as a control variable , And the ball mill level (grinding memory coal), inlet negative pressure and outlet temperature as a controlled amount. Due to the multivariable, strong coupling, large inertia, pure delay, and time-varying model, the ball mill system has led to the traditional single-loop PID control system (ie, the use of coal feed to control the material level, the opening of the cold air door to control the inlet vacuum , the opening of the hot air door to control the outlet temperature) can not be effectively implemented, because the differential pressure control loop and the inlet negative pressure control circuit, the outlet temperature control loop there is a serious coupling, changes in the inlet pressure of the coal mill will affect the change of the cold and hot air volume, This causes a change in the outlet temperature and also affects the differential pressure.
Therefore, using such a common PID single-loop control to forcibly break the relationship between variables, not only does not get a good control effect, but also very easily lead to instability of the boiler combustion system. In addition, the use of coal to control the ball mill inlet and outlet differential pressure, it is difficult to ensure that the ball mill running in the optimization area without blocking phenomenon, so the conventional ball mill load control strategy application effect is not ideal, the system is difficult to put into automatic for a long time.
2.2 Decoupling control literature [1] applied the Bristol-shinskey method to the theoretical analysis of the coupling of the milling system ball mill. The conclusion is that the undecoupled system cannot achieve good control, and the decoupling design is a must.
Literature [1] designed two kinds of compensation methods: feedforward compensation method and triangular matrix method. In [2], the load of the ball mill is represented by a vibration signal, the coal supply is designed as a single loop, and the outlet temperature of the ball mill and the inlet negative pressure loop are based on a multivariable frequency domain design method, and the reverse standardization is given respectively. The (RFN) design method and the inverse Nyquist Array (INA) method are characterized by a simple system structure and easier implementation of the project, as shown in FIG.
For the ball mill inlet negative pressure and outlet temperature, three decoupled control systems were designed in [3]. The practical static decoupling method was determined by experiments to determine the decoupling compensation function, which has certain practicality. See Figure 2.
In [4], on the ball mill inlet negative pressure and outlet temperature control system, the following improvement methods are proposed: designing cross decoupling links on the basis of two single-loop control systems; adding phase compensation in the coal mill outlet temperature control system It can effectively overcome the large inertia of the temperature object and improve the quality of the temperature control system. This improvement has been applied in the Nanjing Thermal Power Plant, see Figure 3.
The ball mill is a nonlinear, slow time-varying complex industrial object. As the perturbation of the parameters leads to the mismatch of the decoupled model, how to construct a decoupling model with self-adaptive capability or a dynamic compensation model becomes the key to the decoupling control. .
2.3 Predictive control The starting point of predictive control is different from the traditional PID control: PID control determines the current control input based on the current and past output measurement and setpoint deviations of the process, while predictive control uses not only current and past differences, Moreover, predictive models are also used to predict future process deviations to determine the current optimal input strategy. Ball mill powder system is a purely delayed industrial process with large inertia, so predictive control can solve this control problem well. In [5], the model algorithm control (MAC) was used to control the ball mill. At the same time, the static error was eliminated by introducing output feedback. The specific MAC control consists of three parts: the internal model of the controlled system, the reference model, and the calculation of control variables. . The reference model is used to generate the reference trajectory, guide the controlled object's output to converge to the given value without overshooting along this desired smooth curve, and use the predictive two-step control algorithm to solve the optimal control scheme for the three loops.
In [6], the problem of solving a large number of linear equations when MAC is decoupled from a multivariable coupling system is proposed, and a new decoupling idea is proposed. The core idea is to assume that Y1, Y2, and Y3 have been decoupled. This is given so that the prediction time domains P1, P2, P3 and the control time domains M1, M2, M3 have different values ​​in different situations.
2.4 Fuzzy Control Fuzzy control is a kind of non-linear control, easy to link with specific production practice experience, solve the control problem of the terminal knowledge model system, has good stability and strong robustness, so it is very suitable for ball mill Complex control object.
For a three-input, three-output, multivariable system with a ball mill, the use of conventional single-step fuzzy control will result in too many rules: if m=7 and n=6, the number of rules is 76=117649, which is for the site. The hardware requirements for control are very high. They can neither achieve the purpose of real-time control nor save the cost of retrofitting. Therefore, fuzzy decoupling control can be used, but also can adopt hierarchical fuzzy control, that is, the first level system is regarded as three independent systems, the calculator controls the role, and the second level is used to adjust the three output levels of the first level. Thus, the total number of first-level control rules is 72×3=147, the total number of second-level control rules is 73=343, and the total number of control rules is 490, which is greatly reduced, as shown in FIG. 4 .
Fuzzy control has a big limitation. Different control algorithms must be adopted for different ball mills. In addition, due to the time-varying characteristics of the ball mill model, the fuzzy control has no good robustness, and the control system responds very well to the perturbation of the controlled object. strong. Therefore, the fuzzy control is often combined with other control algorithms to take full advantage of the advantages of fuzzy control for rapid response and large amplitude of action when the system is in large deviations. In addition to the good static performance of other control algorithms, the ideal control objective is achieved.
2.5 Self-optimizing control Because of the slow time-varying characteristics of the coal mill, the optimal operating point will be affected by various factors. Therefore, to improve the operating economy requires the control system to keep track of changes in the best operating conditions and keep the Good working conditions. From the load characteristic curve of the coal mill, it can be seen that the optimal material level corresponds to the minimum coal consumption. Therefore, considering the direction of reducing the unit consumption of the coal mill as the adjustment objective, the coal consumption of the ball mill is avoided. A difficult to measure parameter. This control method has good robustness. Whether it is the change of the optimal material level caused by changes in air volume, coal type, coal moisture content, or steel ball filling rate, the control method can track the optimal work well. Changes in conditions to achieve the minimum coal consumption. A dynamic stepping search method using predictive comparison can be used to give a coal change signal, then measure the size and direction of changes in coal unit consumption due to signal changes, and then adjust the coal supply according to the required direction and size. See Figure 5.
For self-optimizing control, mill coal unit consumption is often used as a reference standard for control. However, accurate measurement of the amount of powder output is still an industrial problem. If the amount of coal used is used instead of the amount of powder, the time lag and nonlinearity of the system are used. Inevitably lead to larger errors, so self-seeking is not self-optimizing. It can be seen from the running characteristic curve of the ball mill that the change of the vibration of the front bearing bush near the optimal material level is close to zero, so the change of the vibration amount of the front bearing bush can be used as a self-optimizing reference standard to improve the accuracy of the self-optimizing control. Because the wavy line of the lining tile is extended and deformed, the steel ball and the raw coal cannot be raised to a predetermined angle and height, the loading of the steel ball is insufficient, and the proportions of various diameter steel balls are unreasonable, so that grinding of the steel ball and the raw coal cannot reach an ideal state. Both are the influencing factors that increase the useless work and reduce the output. Therefore, the optimal value of self-optimization is only the optimum under the current equipment conditions, but it is not the optimal economic operation of the pulverizing system. With the improvement of the above-mentioned factors affecting the grinding force, this optimal value will be greatly improved. Therefore, the use of high quality and durable linings and a reasonable ratio of ball diameters and ball replenishment strategies combined with self-optimizing control can achieve very good economic benefits. .
2.6 Neural Network Control Since neural networks are characterized by parallel processing, distributed storage, highly fault-tolerant, self-learning capability, strong robustness, and strong adaptability, and they can approach any nonlinear function, they are increasingly attracting attention from the control community. More and more, it is used in complex industrial control processes that are difficult to solve with traditional control. Neural networks do not depend on specific mathematical models, and are very suitable for objects whose dynamic characteristics vary widely with operating conditions. Ball mills, on the other hand, belong to such objects.
The literature [9] uses neuron decoupling controllers for temperature and negative pressure. The neurons use Hebb learning rules and learning rules to combine and respond to changes in the outside world through correlation searches to achieve the self-learning function. And has a good decoupling performance. The literature proposes an inverse system control scheme for ball mill system control and inverse system control based on distributed artificial neural network. A fuzzy working condition division method suitable for distributed control is given and robust PID integration of the inverse system is performed. Compared with the design results by the linear control method, the nonlinear and strong coupling problems of the system can be overcome in a large range, and new methods have been explored for the automatic control of the ball mill in the intermediate storage type pulverizing system that has been difficult to solve. 6.
Third, discuss (1) ball mill level measurement. In the ball mill control optimization program, the material level is an important control parameter, and the direct on-line measurement of the material level has always been a difficult problem in the industrial control field. The traditional measurement method uses the inlet differential pressure to replace the material level. Although the size of the differential pressure and the level of the material level have a certain degree of consistency, but the material level will be affected by a number of factors and parameters, so the traditional measurement The error in the means is so great that the control accuracy does not achieve the desired effect. In recent years, the use of audio or vibration measurement method to measure the ball mill material level has greatly improved the accuracy of the traditional methods, but only with a relevant parameter to identify the ball mill material level is certainly a big error, so you can use Multi-information fusion technology soft measurement means to measure the change of the ball mill's material level. Therefore, many scholars use back-propagation (BP) neural networks to establish a measurement model, that is, to use various quantities related to the level of the ball mill (air volume, pressure difference, inlet/outlet temperature, vibration amount, etc.) as the input to the neural network. Bits are used as outputs to build a measurement model. Of course, training and verification networks require a large amount of sample data (requiring sample data to be comprehensive and accurate), which brings great workload to the experimenter and will also have a relatively large impact on industrial site production. Due to the slow convergence rate of the BP network itself, and the training is likely to fall into a local minimum, the accuracy of the model varies greatly. It is suggested that the wavelet theory be used to deal with the input signal (such as the vibration signal), extract the characteristic parameters and then establish the model, and even use the wavelet neural network instead of the neural network to improve the accuracy of the soft measurement model.
(2) Due to the complexity of the ball mill system itself, the difficulty of modeling is caused, which increases the use of decoupling, prediction and other control methods based on the mathematical model. Since it is difficult to obtain satisfactory control effects due to a fixed single control rule or algorithm, comprehensive consideration must be given to a combination of multiple control methods and their respective advantages. For example, taking into account the characteristics of multivariable, strong coupling, large inertia, pure delay, and dynamic model of time-varying ball milling mechanism powder system, full use of fuzzy control and neural network has a good ability to adapt to nonlinear objects, through and other controls The method is combined to achieve the purpose of controlling and optimizing the ball milling system. In addition, consideration should be given to integrating current intelligent algorithms, such as genetic algorithms, particle swarm optimization, and immune optimization, so as to improve control quality and achieve a safe and economical production mode.
(3) The formulation of the control strategy should fully consider the actual working conditions and characteristics of the ball milling system powder system, and it is not necessary to select those advanced control theories and methods. Considering the ball milling mechanism powder system as an important part of the boiler combustion system, Its disturbance will affect the entire plant's production and operation. Therefore, it is first necessary to ensure the safety and stable operation of the ball milling system powder system, that is, to minimize the unit consumption of milling powder and the overall production efficiency. In actual operation, the system is adjusted to allow a certain range of fluctuations. This point should be taken into consideration when designing the control strategy, so as to reduce the frequent fine-tuning of the volume adjustment in the operation process, which not only ensures the safety of production but also reduces the number of actuators. The frequency of actions is beneficial to improve the stability of the entire operating system. In addition, because the transformation of the ball mill powder system control is carried out on the basis of the original power plant decentralized control system (DCS), the designed control system must be able to be conveniently implemented on the site and construct a practical support system for operation to make full use of it. The existing hardware and software resources of DCS, in particular, further consider the operation support system based on DCS or even field bus control system (FCS).
(4) At present, various types of control systems are implemented and reconstructed differently, and some are based on the original expansion of IPCs and board cards to achieve, some add a small computer network, and some use the original common instruments of the power plant. And instrumentation, just adding its algorithms and functions to achieve. However, due to the complexity of the ball milling system itself, a lot of repetitive work is required for the transformation of different types of ball mills. At present, China's power plants are developing in the direction of large-capacity large-capacity units. The safe, stable, and efficient operation of ball mills is related to the safety and economy of the entire plant's production and operation. Therefore, a universally adaptable ball mill system system control is designed. The plan will have great significance.
The bags could be packed individually for one piece or 10 piece , or just as the quantity you need.To Compare with garbage bags on roll, Garbage bags on sheet could save more space, and you do not need to tear off the bags before usage.
First, the establishment of a mathematical model of the ball mill 1.1 Test Method Modeling Test method modeling is established through the relationship between the input and output signals, for example, the input signal can be used as a step signal or pulse signal to get the system's transfer function. At present, in practice, the differential pressure at the inlet and outlet of the coal mill is generally used to characterize the material level.
1.2 Mechanism Method Modeling Mechanism method modeling is to write various balance equations based on the mechanism of system change: material balance equation, energy balance equation, momentum balance equation, phase balance equation, and reflection of fluid flow, heat transfer, mass transfer, and chemical reaction. The basic equation of the characteristic equation, from which to obtain the desired mathematical model.
Second, the ball mill powder system commonly used control method 2.1 Conventional proportional integral differential (PlD) control ball mill system is a three-input three-output control object, coal supply, hot air door opening and recirculation (cold) throttle opening as a control variable , And the ball mill level (grinding memory coal), inlet negative pressure and outlet temperature as a controlled amount. Due to the multivariable, strong coupling, large inertia, pure delay, and time-varying model, the ball mill system has led to the traditional single-loop PID control system (ie, the use of coal feed to control the material level, the opening of the cold air door to control the inlet vacuum , the opening of the hot air door to control the outlet temperature) can not be effectively implemented, because the differential pressure control loop and the inlet negative pressure control circuit, the outlet temperature control loop there is a serious coupling, changes in the inlet pressure of the coal mill will affect the change of the cold and hot air volume, This causes a change in the outlet temperature and also affects the differential pressure.
Therefore, using such a common PID single-loop control to forcibly break the relationship between variables, not only does not get a good control effect, but also very easily lead to instability of the boiler combustion system. In addition, the use of coal to control the ball mill inlet and outlet differential pressure, it is difficult to ensure that the ball mill running in the optimization area without blocking phenomenon, so the conventional ball mill load control strategy application effect is not ideal, the system is difficult to put into automatic for a long time.
2.2 Decoupling control literature [1] applied the Bristol-shinskey method to the theoretical analysis of the coupling of the milling system ball mill. The conclusion is that the undecoupled system cannot achieve good control, and the decoupling design is a must.
Literature [1] designed two kinds of compensation methods: feedforward compensation method and triangular matrix method. In [2], the load of the ball mill is represented by a vibration signal, the coal supply is designed as a single loop, and the outlet temperature of the ball mill and the inlet negative pressure loop are based on a multivariable frequency domain design method, and the reverse standardization is given respectively. The (RFN) design method and the inverse Nyquist Array (INA) method are characterized by a simple system structure and easier implementation of the project, as shown in FIG.
For the ball mill inlet negative pressure and outlet temperature, three decoupled control systems were designed in [3]. The practical static decoupling method was determined by experiments to determine the decoupling compensation function, which has certain practicality. See Figure 2.
In [4], on the ball mill inlet negative pressure and outlet temperature control system, the following improvement methods are proposed: designing cross decoupling links on the basis of two single-loop control systems; adding phase compensation in the coal mill outlet temperature control system It can effectively overcome the large inertia of the temperature object and improve the quality of the temperature control system. This improvement has been applied in the Nanjing Thermal Power Plant, see Figure 3.
The ball mill is a nonlinear, slow time-varying complex industrial object. As the perturbation of the parameters leads to the mismatch of the decoupled model, how to construct a decoupling model with self-adaptive capability or a dynamic compensation model becomes the key to the decoupling control. .
2.3 Predictive control The starting point of predictive control is different from the traditional PID control: PID control determines the current control input based on the current and past output measurement and setpoint deviations of the process, while predictive control uses not only current and past differences, Moreover, predictive models are also used to predict future process deviations to determine the current optimal input strategy. Ball mill powder system is a purely delayed industrial process with large inertia, so predictive control can solve this control problem well. In [5], the model algorithm control (MAC) was used to control the ball mill. At the same time, the static error was eliminated by introducing output feedback. The specific MAC control consists of three parts: the internal model of the controlled system, the reference model, and the calculation of control variables. . The reference model is used to generate the reference trajectory, guide the controlled object's output to converge to the given value without overshooting along this desired smooth curve, and use the predictive two-step control algorithm to solve the optimal control scheme for the three loops.
In [6], the problem of solving a large number of linear equations when MAC is decoupled from a multivariable coupling system is proposed, and a new decoupling idea is proposed. The core idea is to assume that Y1, Y2, and Y3 have been decoupled. This is given so that the prediction time domains P1, P2, P3 and the control time domains M1, M2, M3 have different values ​​in different situations.
2.4 Fuzzy Control Fuzzy control is a kind of non-linear control, easy to link with specific production practice experience, solve the control problem of the terminal knowledge model system, has good stability and strong robustness, so it is very suitable for ball mill Complex control object.
For a three-input, three-output, multivariable system with a ball mill, the use of conventional single-step fuzzy control will result in too many rules: if m=7 and n=6, the number of rules is 76=117649, which is for the site. The hardware requirements for control are very high. They can neither achieve the purpose of real-time control nor save the cost of retrofitting. Therefore, fuzzy decoupling control can be used, but also can adopt hierarchical fuzzy control, that is, the first level system is regarded as three independent systems, the calculator controls the role, and the second level is used to adjust the three output levels of the first level. Thus, the total number of first-level control rules is 72×3=147, the total number of second-level control rules is 73=343, and the total number of control rules is 490, which is greatly reduced, as shown in FIG. 4 .
Fuzzy control has a big limitation. Different control algorithms must be adopted for different ball mills. In addition, due to the time-varying characteristics of the ball mill model, the fuzzy control has no good robustness, and the control system responds very well to the perturbation of the controlled object. strong. Therefore, the fuzzy control is often combined with other control algorithms to take full advantage of the advantages of fuzzy control for rapid response and large amplitude of action when the system is in large deviations. In addition to the good static performance of other control algorithms, the ideal control objective is achieved.
2.5 Self-optimizing control Because of the slow time-varying characteristics of the coal mill, the optimal operating point will be affected by various factors. Therefore, to improve the operating economy requires the control system to keep track of changes in the best operating conditions and keep the Good working conditions. From the load characteristic curve of the coal mill, it can be seen that the optimal material level corresponds to the minimum coal consumption. Therefore, considering the direction of reducing the unit consumption of the coal mill as the adjustment objective, the coal consumption of the ball mill is avoided. A difficult to measure parameter. This control method has good robustness. Whether it is the change of the optimal material level caused by changes in air volume, coal type, coal moisture content, or steel ball filling rate, the control method can track the optimal work well. Changes in conditions to achieve the minimum coal consumption. A dynamic stepping search method using predictive comparison can be used to give a coal change signal, then measure the size and direction of changes in coal unit consumption due to signal changes, and then adjust the coal supply according to the required direction and size. See Figure 5.
For self-optimizing control, mill coal unit consumption is often used as a reference standard for control. However, accurate measurement of the amount of powder output is still an industrial problem. If the amount of coal used is used instead of the amount of powder, the time lag and nonlinearity of the system are used. Inevitably lead to larger errors, so self-seeking is not self-optimizing. It can be seen from the running characteristic curve of the ball mill that the change of the vibration of the front bearing bush near the optimal material level is close to zero, so the change of the vibration amount of the front bearing bush can be used as a self-optimizing reference standard to improve the accuracy of the self-optimizing control. Because the wavy line of the lining tile is extended and deformed, the steel ball and the raw coal cannot be raised to a predetermined angle and height, the loading of the steel ball is insufficient, and the proportions of various diameter steel balls are unreasonable, so that grinding of the steel ball and the raw coal cannot reach an ideal state. Both are the influencing factors that increase the useless work and reduce the output. Therefore, the optimal value of self-optimization is only the optimum under the current equipment conditions, but it is not the optimal economic operation of the pulverizing system. With the improvement of the above-mentioned factors affecting the grinding force, this optimal value will be greatly improved. Therefore, the use of high quality and durable linings and a reasonable ratio of ball diameters and ball replenishment strategies combined with self-optimizing control can achieve very good economic benefits. .
2.6 Neural Network Control Since neural networks are characterized by parallel processing, distributed storage, highly fault-tolerant, self-learning capability, strong robustness, and strong adaptability, and they can approach any nonlinear function, they are increasingly attracting attention from the control community. More and more, it is used in complex industrial control processes that are difficult to solve with traditional control. Neural networks do not depend on specific mathematical models, and are very suitable for objects whose dynamic characteristics vary widely with operating conditions. Ball mills, on the other hand, belong to such objects.
The literature [9] uses neuron decoupling controllers for temperature and negative pressure. The neurons use Hebb learning rules and learning rules to combine and respond to changes in the outside world through correlation searches to achieve the self-learning function. And has a good decoupling performance. The literature proposes an inverse system control scheme for ball mill system control and inverse system control based on distributed artificial neural network. A fuzzy working condition division method suitable for distributed control is given and robust PID integration of the inverse system is performed. Compared with the design results by the linear control method, the nonlinear and strong coupling problems of the system can be overcome in a large range, and new methods have been explored for the automatic control of the ball mill in the intermediate storage type pulverizing system that has been difficult to solve. 6.
Third, discuss (1) ball mill level measurement. In the ball mill control optimization program, the material level is an important control parameter, and the direct on-line measurement of the material level has always been a difficult problem in the industrial control field. The traditional measurement method uses the inlet differential pressure to replace the material level. Although the size of the differential pressure and the level of the material level have a certain degree of consistency, but the material level will be affected by a number of factors and parameters, so the traditional measurement The error in the means is so great that the control accuracy does not achieve the desired effect. In recent years, the use of audio or vibration measurement method to measure the ball mill material level has greatly improved the accuracy of the traditional methods, but only with a relevant parameter to identify the ball mill material level is certainly a big error, so you can use Multi-information fusion technology soft measurement means to measure the change of the ball mill's material level. Therefore, many scholars use back-propagation (BP) neural networks to establish a measurement model, that is, to use various quantities related to the level of the ball mill (air volume, pressure difference, inlet/outlet temperature, vibration amount, etc.) as the input to the neural network. Bits are used as outputs to build a measurement model. Of course, training and verification networks require a large amount of sample data (requiring sample data to be comprehensive and accurate), which brings great workload to the experimenter and will also have a relatively large impact on industrial site production. Due to the slow convergence rate of the BP network itself, and the training is likely to fall into a local minimum, the accuracy of the model varies greatly. It is suggested that the wavelet theory be used to deal with the input signal (such as the vibration signal), extract the characteristic parameters and then establish the model, and even use the wavelet neural network instead of the neural network to improve the accuracy of the soft measurement model.
(2) Due to the complexity of the ball mill system itself, the difficulty of modeling is caused, which increases the use of decoupling, prediction and other control methods based on the mathematical model. Since it is difficult to obtain satisfactory control effects due to a fixed single control rule or algorithm, comprehensive consideration must be given to a combination of multiple control methods and their respective advantages. For example, taking into account the characteristics of multivariable, strong coupling, large inertia, pure delay, and dynamic model of time-varying ball milling mechanism powder system, full use of fuzzy control and neural network has a good ability to adapt to nonlinear objects, through and other controls The method is combined to achieve the purpose of controlling and optimizing the ball milling system. In addition, consideration should be given to integrating current intelligent algorithms, such as genetic algorithms, particle swarm optimization, and immune optimization, so as to improve control quality and achieve a safe and economical production mode.
(3) The formulation of the control strategy should fully consider the actual working conditions and characteristics of the ball milling system powder system, and it is not necessary to select those advanced control theories and methods. Considering the ball milling mechanism powder system as an important part of the boiler combustion system, Its disturbance will affect the entire plant's production and operation. Therefore, it is first necessary to ensure the safety and stable operation of the ball milling system powder system, that is, to minimize the unit consumption of milling powder and the overall production efficiency. In actual operation, the system is adjusted to allow a certain range of fluctuations. This point should be taken into consideration when designing the control strategy, so as to reduce the frequent fine-tuning of the volume adjustment in the operation process, which not only ensures the safety of production but also reduces the number of actuators. The frequency of actions is beneficial to improve the stability of the entire operating system. In addition, because the transformation of the ball mill powder system control is carried out on the basis of the original power plant decentralized control system (DCS), the designed control system must be able to be conveniently implemented on the site and construct a practical support system for operation to make full use of it. The existing hardware and software resources of DCS, in particular, further consider the operation support system based on DCS or even field bus control system (FCS).
(4) At present, various types of control systems are implemented and reconstructed differently, and some are based on the original expansion of IPCs and board cards to achieve, some add a small computer network, and some use the original common instruments of the power plant. And instrumentation, just adding its algorithms and functions to achieve. However, due to the complexity of the ball milling system itself, a lot of repetitive work is required for the transformation of different types of ball mills. At present, China's power plants are developing in the direction of large-capacity large-capacity units. The safe, stable, and efficient operation of ball mills is related to the safety and economy of the entire plant's production and operation. Therefore, a universally adaptable ball mill system system control is designed. The plan will have great significance.
Suitable for kitchen, office, courtyard, lawn and also the dustbin.
The garbage bag has large capacity, reliable quality, can carry a large amount of garbage, environmental protection and durability.The bags could be packed individually for one piece or 10 piece , or just as the quantity you need.To Compare with garbage bags on roll, Garbage bags on sheet could save more space, and you do not need to tear off the bags before usage.
Bin Liners, Small Trash Bags, Kitchen Trash Bags, Clear Garbage Bags, Black Trash Bag, Refused Sack
BILLION PLASTIC MANUFACTURING CO.,LTD, JIANGMEN , https://www.billion-plasticsjm.com